The present invention relates to the field of digital media distribution. More specifically, the present invention relates to a high quality video encoding termed a “Home-Video Digital-Master Package” (HVP).
Video and audio content, such as movies and television shows, are beginning to become widely available for viewing on a variety of digital media devices. The HVP allows video content source providers to encode video data into a variety of formats for a variety of devices with reduced artifacts. Digital media devices include computers, set-top video boxes, digital video recorders, mobile phones, personal digital assistants, DVD players, and handheld or portable video players (e.g. iPhone, Zune, ipod).
Digital distribution uses data communications networks, such as the Internet, local area networks, local or wide-area terrestrial or satellite wireless networks, cellular data networks, and other open or proprietary networks to distribute content to digital media devices. Digital media devices can download content via a data communications networks on demand during viewing, referred to as streaming, and/or for storage in advance of later viewings.
Currently, there is a large amount of content already digitally mastered for distribution via DVD, digital broadcasting, or more recently, digital television broadcasting. However, for such data to be prepared for digital media devices, the digitally mastered content must typically be further processed and encoded. The file sizes for digitally-mastered content are often very large compared to the bandwidth available on data communications networks. Further, the file size also typically exceeds the data storage capacities of digital media devices. Accordingly, digitally-mastered content is often compressed for digital distribution purposes. Digital media devices typically use any standard or proprietary data compression known in the art, e.g. MPEG-2, DiVX, or the like.
Digital media devices, especially portable and handheld devices, often have relatively low resolution displays compared with the resolution of the digitally mastered content. Furthermore, the pixel aspect ratios of digital media devices can differ from the pixel aspect ratio used to digitally master content. For example, many computers and other digital media devices use square pixels (e.g. 1:1), while DVD format video typically have a non-square pixel aspect ratio of 0.9:1.0. Thus, digitally mastered content must often be resized to account for differences in resolution and pixel aspect ratio between the digital master version of the content and the display formats of digital media devices.
Due to the variety of different standard formats used for generating and distributing digital master versions of content and the plethora of different resolutions, pixel aspect ratios, and compression formats required by digital media devices, the content encoded for digital media devices by current digital distribution techniques is often low quality. In particular, digital distribution systems may compress, decompress, and recompress content multiple times using different data compression schemes. Because typical digital distribution systems and digital media devices typically rely on lossy data compression schemes; quality is reduced and compression artifacts are introduce every time the content is compressed. By compressing content multiple times, digital distribution systems often exacerbate compression artifacts and substantially reduce the quality of content for digital media devices.
As merely an example, the inventor of the present invention has studied the quality of the video being displayed on digital media devices. Many artifacts of such video are determined to be caused by one or more of the intermediate conversion steps between film format to hand-held device format. In a specific example, a traditional acetate film media is scanned with a film scanner (e.g. telecine) to form digital video data (e.g. 720×480 pixels); next, the digital video data is typically encoded into MPEG2 format (e.g. for DVDs) (e.g. 720×480 pixels); the MPEG2 encoded DVD resolution data is then provided to a video content providing service (e.g. iTunes). Subsequently, the video content providing service first decodes the MPEG2 format video; then resizes the 720×480 pixel resolution video to the target resolution (e.g. 640×480, 320×480, etc.); and then recompresses the video in the desired encoding scheme (e.g. H.264 for iTunes).
FIGS. 8A-C illustrate a prior art problem. More specifically, FIGS. 8A-C illustrate the problem of multiple resizing of video content upon a source image. FIG. 8A illustrates source image having a first image resolution 800 to be resized to a second image resolution 810. Because of differences in resolution, a resulting image 820 will often have a number of “beat frequencies” shown as dark lines in resulting image 820. In one example, in FIG. 8A, the source image may be a be a high resolution image, e.g. 1960×1080 resolution, and second resolution 810 may be 720×480 resolution. Next, FIG. 8B illustrates resulting image 820 being resized to a third image resolution 830. Because of additional differences in resolution, a resulting image 840 will often have an additional number of “beat frequencies” shown as additional dark lines in resulting image 840. In one example, in FIG. 8B, the third image resolution 810 may be 640×480. In contrast, in FIG. 8C, the inventors of the present invention recognize that direct resizing from the source image having first resolution 800 to third image resolution 830 results in resulting image 850 with reduced “beat frequencies.” In this example, resulting image 850 includes fewer dark lines compared to resulting image 840.
Additionally, in one specific example, because MPEG2 encoding for the DVD resolution image uses blocks of 8×8 pixels, the edges of many of the encoding blocks become visible as vertical and horizontal lines when resizing the video images to 640×480. Another problem is that the video images are encoded and decoded multiple times, causing a degradation in the colors. In the example above, the digital video data is encoded using MPEG2, is decoded, and then re-encoded to H.264.
FIGS. 7A-D illustrate another prior art drawback. More specifically, FIGS. 7A-C illustrate a problem with typical block-based, or shift-intolerant encoding or compression schemes. FIG. 7A illustrates an image 700 including data to be compressed by a block-based compression scheme, such as MPEG2. In FIG. 7A, each group of four pixels is used to determine a pixel value in a compressed image. In the example in FIG. 7B, an image 710 illustrates the resulting compressed image reproducing the checker board-type pattern.
In FIG. 7C, image 720 illustrates image 700, however, the grouping of four pixels used to determine a pixel value is shifted vertically and horizontally by one pixel. An example of where this may happen is when scanning film media. It is virtually impossible to guarantee scanning at the same position of each image because of film printing and duplication registration errors, scanning misalignments, or the like. In light of this, scanned film images often are shifted for different images as shown in FIGS. 7A and 7C.
In the example in FIG. 7C an image 730 illustrates the compressed image determined from image 720 in FIG. 7C. As can be seen, the data values in image 730 are uniformly 0.5, in contrast, the data values in image 710 (FIG. 7B) vary from 1 to 0 in a checkerboard-type pattern. Accordingly, if a viewer were to view image 710 and 730 in succession, the viewer would see the output images flicker and go to gray on the screen due simply to the compression scheme being shift-intolerant, e.g. block-based.
It is therefore desirable for a system and method to provide the highest-quality content possible for digital media devices regardless of their display requirements. It is further desirable for a system and method to simplify the process of converting content for digital media devices. It is also desirable for a system and method to integrate easily with both legacy content sources, including film-based content sources, and more recent digital sources. Further, the inventors desire a method for providing service providers with video content that will have fewer visual artifacts when resized and/or encoded in a desired format.